Speech Bandwidth Extension With Short-Time High-Band Post-Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing bandwidth extension technologies in the time domain suffer from deficient performance indicators, resulting in restored speech signals with rustling and unclear sound due to inaccuracies in high frequency signal restoration.
Innovation Solution
Implement short-time post-filtering processing on synthesized high band signals to obtain a short-time filtered signal, calculating high frequency gain based on this filtered signal to improve encoding and decoding performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional bandwidth extension technology is used in time domain, then encoding can be performed on wider bandwidth signals with unchanged bit rate, but the restored speech signal quality deteriorates with rustling and unclear sound
Solution Approach 1:
The patent applies post-filtering processing to the synthesized high band signal before calculating the high frequency gain. This preliminary action of filtering removes unwanted spectral components and aligns the spectral envelope closer to the original signal, preventing the rustling artifact from occurring in the first place during the synthesis process
Solution Approach 2:
The patent introduces a feedback mechanism where the synthesized high band signal is filtered and compared with the original high band signal to calculate an adjusted high frequency gain. This feedback loop continuously optimizes the gain parameter to minimize the difference between synthesized and original signals, thereby improving restored speech quality while maintaining bandwidth extension
2Device complexity
If high frequency gain is calculated directly from synthesized high band signal without filtering, then encoding complexity is reduced, but spectral envelope alignment with original signal deteriorates
Solution Approach 1:
The patent applies a selective post-filtering operation only to the synthesized high band signal before gain calculation, rather than processing the entire signal chain. This partial action approach achieves the necessary spectral envelope alignment improvement without adding excessive complexity to the overall encoding system
Solution Approach 2:
The patent modifies the high frequency gain parameter by calculating it from the filtered synthesized signal rather than the raw synthesized signal. This parameter change transforms the gain calculation to account for the spectral shaping introduced by post-filtering, thereby achieving better spectral envelope alignment with the original signal
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Embodiments of the present invention provide an encoding method, a decoding method, an encoding apparatus, a decoding apparatus, a transmitter, a receiver, and a communications system. The encoding method includes: dividing a to-be-encoded time-domain signal into a low band signal and a high band signal; performing encoding on the low band signal to obtain a low frequency encoding parameter; performing encoding on the high band signal to obtain a high frequency encoding parameter, and obtaining a synthesized high band signal according to the low frequency encoding parameter and the high frequency encoding parameter; performing short-time post-filtering processing on the synthesized high band signal to obtain a short-time filtering signal, where, compared with a shape of a spectral envelope of the synthesized high band signal, a shape of a spectral envelope of the short-time filtering signal is closer to a shape of a spectral envelope of the high band signal; and calculating a high frequency gain based on the high band signal and the short-time filtering signal. A technical solution according to the embodiments of the present invention can improve an encoding and/or decoding effect.